Synthesis of ethyl octyl ether from diethyl carbonate and 1-octanol over solid catalysts. A screening study

Article abstract of DOI:10.1016/j.apcata.2011.10.037

The synthesis of ethyl octyl ether (EOE) from a mixture of diethyl carbonate (DEC) and 1-octanol (1:2 molar ratio) over several solid catalysts was studied in batch mode at 150 °C and 25 bar. Catalyst screening revealed that EOE could be successfully obtained over some acid catalysts. In particular the highest yield was achieved over acid ion-exchange resins (33% after 8 h). A reaction scheme of the process is proposed. Selectivity to EOE was mainly affected by the production of diethyl ether (DEE) and di-n-octyl ether (DNOE). However, EOE was the main ether obtained (60 mol%), followed by DEE (20 mol%) and DNOE (20 mol%). By comparing the behavior of several acid resins, it was seen that the synthesis of EOE was highly related to the structural resin properties. It was found that the accessibility of DEC and 1-octanol to acid centers was improved over highly swollen and low polymer density resins. Thus, gel-type resins with low divinylbenzene content are the most suitable to produce EOE (e.g., Amberlyst 121, Dowex 50Wx2-100 and CT224).

Full text of DOI:10.1016/j.apcata.2011.10.037

Applied Catalysis A: General 413–414 (2012) 21–29
Contents lists available at SciVerse ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Synthesis of ethyl octyl ether from diethyl carbonate and 1-octanol over solid
catalysts. A screening study
∗
J. Guilera, R. Bringué, E. Ramírez, M. Iborra, J. Tejero
Chemical Engineering Department, Faculty of Chemistry, University of Barcelona, C/Martí i Franquès 1, 08028, Barcelona, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 August 2011
Received in revised form 7 October 2011
Accepted 24 October 2011
Available online 6 November 2011
The synthesis of ethyl octyl ether (EOE) from a mixture of diethyl carbonate (DEC) and 1-octanol (1:2
molar ratio) over several solid catalysts was studied in batch mode at 150 C and 25 bar. Catalyst screening
revealed that EOE could be successfully obtained over some acid catalysts. In particular the highest yield
was achieved over acid ion-exchange resins (33% after 8 h). A reaction scheme of the process is proposed.
Selectivity to EOE was mainly affected by the production of diethyl ether (DEE) and di-n-octyl ether
◦
(
(
DNOE). However, EOE was the main ether obtained (60 mol%), followed by DEE (20 mol%) and DNOE
20 mol%). By comparing the behavior of several acid resins, it was seen that the synthesis of EOE was
Keywords:
Ethyl octyl ether
Diethyl carbonate
Ion-exchange resins
Diesel reformulation
Oxygenated compounds
highly related to the structural resin properties. It was found that the accessibility of DEC and 1-octanol
to acid centers was improved over highly swollen and low polymer density resins. Thus, gel-type resins
with low divinylbenzene content are the most suitable to produce EOE (e.g., Amberlyst 121, Dowex®
50Wx2-100 and CT224).
©
2011 Elsevier B.V. All rights reserved.
1
. Introduction
A green alkylation route to produce asymmetrical ethers is
achieved with carbonates. Dimethyl carbonate has been proposed
as methylating agent of several substances and reacts either as
a methoxycarbonylating or as a methylating agent depending on
the operation conditions [13]. In particular, the octanol alkylation
from dimethyl carbonate clearly showed to be more efficient than
using directly methanol [14]. As dimethyl carbonate, diethyl car-
bonate (DEC) is generally accepted as an environmentally benign
ethylating agent [15–19]. An advantage of using DEC is that it can
be obtained from ethanol (Eq. (1)). As a consequence, EOE would
be a synthetic bio-fuel and could get the proper tax reduction,
compensating partially their production costs higher than current
commercial diesel.
In order to reduce the environmental impact of diesel exhaust
emissions (particulate matter, unburned hydrocarbons, CO and
NOx), it is mandatory to improve the burning quality of diesel fuels
[
1,2]. The ideal diesel fuel for accomplishing this aim should be
composed by linear paraffins with 10–20 carbon atoms, high cetane
number and proper cold flow properties [3]. As a result, oil industry
addresses its efforts to increase the yield in diesel cuts of refin-
ery processes and thus commercialize better diesel fuel blends.
A complementary strategy could be reformulating diesel fuels by
including linear ethers [4–7]. Ethyl octyl ether (EOE) is an interest-
ing candidate to be used in blends with commercial gasoil because
◦
of its properties: 10% oxygen content (w/w), 187 C boiling point,
2
4
0
d
of 0.771, cetane number of 97 and satisfactory lubricity [8].
Industrially, it has been claimed that linear long chain ethers
combined with peroxides as supplements to diesel fuels provide
an environmental reduction of air pollutants. An amount of ethers
about 1–5% (v/v) increased the cetane number from 2 to 20 points
(1)
[
1]. These ethers can be prepared by the traditional Williamson
A drawback to use alkyl carbonates is that their decomposition
generates CO2 as a by-product. However, the formation of carbon-
ates from CO2 is an interesting way for recycling it to fuels. Several
advances in this direction have been reported and cyclic carbonate
synthesis is already been industrialized [20–22]. Focused on lin-
ear carbonates, CO2 reacts with alcohols in the presence of metal
method from an alkoxide and alkyl halide [9]. However, dialkyl
ethers are readily produced by acid catalyzed dehydration of alco-
hols more efficiently [10–12].
complexes (Eq. (2)). Due to the problems with the hydrolysis of
the carbonate, 3 A˚ molecular sieves were used as drying agents
to extract out the formed water. By using a dehydrative agent, an
∗ Corresponding author. Tel.: +34 93 402 1308; fax: +34 93 402 1291.
E-mail address: jtejero@ub.edu (J. Tejero).
0
926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.10.037

2
2
J. Guilera et al. / Applied Catalysis A: General 413–414 (2012) 21–29
Nomenclature
DEC
DEE
DNOE
DOC
dp
DVB
EOC
EOE
EtOH
diethyl carbonate
diethyl ether
di-n-octyl ether
dioctyl carbonate
particle diameter (mm)
divinylbenzene
ethyl octyl carbonate
ethyl octyl ether
ethanol
G
gel-type resin
acid capacity (meq/g)
+
[
H ]
+
[
H ]/Vsp acid centers per volume unit in a swollen polymer
+
3
state (meq H /cm )
ISEC
M
inverse steric exclusion chromatography
macroreticular resin
OcOH
1-octanol
basic capacity (meq/g)
−
[
OH ]
Fig. 1. ISEC pattern for used macroreticular resins.
RDEC/1-octanol DEC to 1-octanol initial molar ratio (mol/mol)
2
SBET
BET surface area in dry state (m /g)
Amberlyst 70 and Amberlyst 121), Purolite (CT124 and CT224),
Sigma (Dowex 50Wx8-100, Dowex 50Wx2-100, Nafion® NR 50,
EOE
S
selectivity to EOE with respect to DEC (%, mol/mol)
DEC
◦
T
temperature ( C)
acid ␥-Al O , basic ␥-Al O ) and Südchemie (H-BEA-25).
2
3
2
3
◦
Tmax
Vsp
W
maximum operating temperature ( C)
volume of the swollen polymer phase (cm /g)
A wide range of styrene-DVB resins were selected. In particular,
a weak and a strong basic resins used in industrial processing were
tested to check the behavior of basic catalysts (Table 1) [25]. Besides
basic resins, 14 acid resins with different structural properties were
tested (Table 2). Resins with low divinylbenzene (DVB) content are
less crosslinked, and as a consequence with a more flexible struc-
ture. In the present work, the polarity of the reaction medium was
high since 1-octanol was the main reactant and ethanol and water
were obtained in significant amounts. Since ion-exchange resins
swell in polar media, their morphology changes in the course of the
reaction and non-permanent pores appear. A useful description of
nature and characteristics of these pores in aqueous phase can be
obtained from inverse steric exclusion chromatography (ISEC) data
3
mass of dry catalyst (g)
DEC conversion (%, mol/mol)
X
DEC
EOE
Y
yield to EOE with respect to DEC (%, mol/mol)
DEC
interesting dimethyl carbonate yield was achieved (55% based on
MeOH) and byproducts were not significantly produced [23].
(
2)
Recently, EOE was synthesized successfully from ethanol and
-octanol over acid gel-type resins at mild conditions (T = 150 C,
[
12,26,27]. In macroreticular structures a part of new open spaces in
◦
1
the range of mesopores can be characterized by the cylindrical pore
model. However, this model is not applicable to describe spaces
between polymer chains formed as a result of polymer swelling. A
good view of the three-dimensional network of swollen polymer
is given by the geometrical model developed by Ogston [28], in
which micropores are described by spaces between randomly ori-
ented rigid rods. The characteristic parameter of this model is the
specific volume of the swollen polymer (volume of the free space
plus that occupied by the skeleton), Vsp. The Ogston model also
allows to distinguish zones of swollen gel phase of different den-
sity or polymer chain concentration (total rod length per volume
P = 20 bar) [24]. In addition to EOE, diethyl ether (DEE) and di-n-
octyl ether (DNOE) were detected as by-products. The differences
between these ethers as diesel components are significant: DNOE
has excellent properties to be blended with commercial diesel,
whereas the high volatility of DEE is a serious drawback for its addi-
tion in large quantities to the diesel pool. Because acid resins easily
catalyze the synthesis of short chain ethers, ethanol is preferred to
be the limiting reactant in order to reduce DEE formation and to
increase EOE production.
The present paper is devoted to study the liquid phase synthe-
sis of EOE from DEC and 1-octanol over solid catalysts. A catalyst
screening is carried out in order to select suitable catalysts for
obtaining EOE. Moreover, the effect of the catalyst structure is
related to DEC conversion and EOE yield.
−2
unit of swollen polymer, nm ) [26,29,30].
Distribution of the different polymer density zones of the
swollen resins in water is shown in Figs.
1 (macroreticu-
lar) and 2 (gel-type). As seen, when the DVB resin content is high
the polymer density is accordingly high, and Vsp values small. For
instance, high crosslinked macroreticular resins such as Amberlyst
2
. Experimental
−
2
1
5 have a very high polymer density (1.5–2 nm ) and Vsp of about
3
2
.1. Materials
0.8 cm /g, being as a result poorly accessible polymer. On the con-
trary, a very low-crosslinked gel resin such as Amberlyst 121 has
much more space between chains (density: 0.2–0.8 nm ) and Vsp
−
2
1
-Octanol (≥99.5%, Fluka) and DEC (≥98%, Fluka) were used as
3
reactants. Distilled water, ethanol (≥99.8%, Panreac), DEE (≥99%,
Panreac), 1-octene (≥97%, Fluka) and DNOE (≥97%, Fluka) were
used for analysis purposes. EOE was obtained and purified by rec-
tification to 99% (w/w) in our lab.
is about 3.3 cm /g. The pore distribution of the swollen gel phase in
water is probably quite representative of the morphology of the cat-
alyst in the reaction medium because the resin swelling in alcohols
and water is comparable [12,26]. Thus, active centers in swollen
state of low crosslinked gel-type resins are by far more accessible,
and as a consequence the number of acid centers that could take
part in the reaction is far higher.
Tested catalyst were supplied by Rohm and Haas (Amberlyst
6 OH, Amberlyst 21, Amberlyst 15, Amberlyst 35, Amberlyst
8, Amberlyst 46, Amberlyst 16, Amberlyst 36, Amberlyst 39,
2
4

J. Guilera et al. / Applied Catalysis A: General 413–414 (2012) 21–29
23
Table 1
Properties of tested basic resins.
Typeb
Basicity (meq OH /g)
−
SBETa (m /g)
2
dp (mm)
Tmax ( C)
◦
Catalyst
Short name
Basic strength
Amberlyst 21
Amberlyst 26 OH
A 21
A 26
Weak
Strong
M
M
4.6
25
30
0.58
0.63
100
60
2–2.6 [43,44]
a
2
BET surface area in dry state (N2 for SBET ≥ 1 m /g).
b
M = macroreticular.
Table 2
Properties and structural parameters of tested acid resins.
Typeb
DVB %
Acidity (meq H /g)
+
Vsp (cm /g)
3
SBETa (m /g)
2
dp (mm)
Tmax ( C)
◦
Catalyst
Short name
Amberlyst 15
Amberlyst 35
Amberlyst 48
Amberlyst 46
Amberlyst 16
Amberlyst 36
Amberlyst 39
A 15
A 35
A 48
A 46
A 16
A 36
A 39
A 70
Nafion 50
A 121
Dow 8
Dow 2
CT124
CT224
M
M
M
M
M
M
M
M
G
G
G
G
G
20
20
4.81
4.8
5.62
0.43
4.8
0.823
0.736
0.563
0.16
1.245
0.999
1.451
1.189
42.01
28.9
33.8
57.4
1.69
21.0
0.09
0.02
0.35
0.02
0.23
1.32
0.07
0.945
0.65
0.66
120
150
High
High
12 [45]
12 [46]
8 [46]
8
c
0.78
0.56
0.57
0.54
0.55
2.40
0.44
0.17
0.25
0.76
0.33
120
130
150
130
190
180
130
150
150
5.4
5.0
Amberlyst 70d
Nafion NR-50
Amberlyst 121
Dowex 50Wx8-100
Dowex 50Wx2-100
CT124
3.01
0.89
4.8
4.83
4.83
5.0
2 [47]
3.263
1.627
2.655
1.994
1.811
8
2
4
4
CT224
G
5.34
150
a
2
2
BET surface area in dry state (N2 for SBET ≥ 1 m /g; Kr for SBET < 1 m /g) – Accusorb ASAP 2020, Micromeritics-, samples were dried at vacuum.
M = macroreticular; G = gel-type.
Sulfonated only at the polymer surface.
b
c
d
Chlorine substituting hydrogen atoms in polymer chains.
On the other hand, tested resins shows a wide range of acid
In addition to ion exchange resins, a zeolite and two alumi-
capacities. Conventional sulfonated resins (e.g., Amberlyst 15) have
a concentration of about a sulfonic group per styrene ring, sta-
tistically as para-substitution [31,32]. Over-sulfonated resins (e.g.,
Amberlyst 35) have a higher concentration of sulfonic groups
because they have in some styrene rings more than one sul-
fonic group. The additional sulfonic groups are predominantly
distributed close to the particle surface [33]. Amberlyst 46 is a
resin sulfonated only at the polymer surface. It has all their acid
sites accessible but its acid capacity is about a tenth of the other.
Amberlyst 70 should be highlighted. It has chlorine atoms in its
structure, which confer this catalyst a higher thermal stability, but
nas (␥-Al O ) were tested (Table 3). Both aluminas were supplied
2
3
as activated form. Textural characteristics were derived from N2
adsorption–desorption isotherms recorded at 77 K with an Accu-
sorb ASAP 2020 sorptometer from Micromeritics (Norcross, G.A.),
◦
after outgassing the samples for 4 h at 500 C. The SiO /Al O molar
2
2
3
ratio of H-BEA-25 was measured by X-ray fluorescence (PW1400,
Philips; detector: LiF crystal; excitation source: Rh). Finally, skele-
tal density was measured by He displacement in an Accupyc 1330
(Micromeritics).
2.2. Synthesis
+
its acid capacity is only about 3 meq H /g. Finally, NR50 (a gel-type
®
copolymer of Teflon and perfluoro-alkanesulfonic monomers)
2.2.1. Apparatus and analysis
was also tested. Fluorine atoms of polymer chains upgrade ther-
mal stability of NR50 and give a higher acid strength than PS-DVB
resins.
Experiments were carried out in a 100 mL-stainless steel auto-
clave operated in batch mode. Temperature was controlled by
an electrical furnace. One of the outlets of the reactor was con-
nected directly to a liquid sampling valve, which injected 0.1 ␮L
of pressurized liquid into a gas–liquid chromatograph. The liquid
composition was analyzed hourly by using a split mode opera-
tion in a HP6890A GLC apparatus equipped with a TCD detector. A
5
0 m × 0.2 mm × 0.5 ␮m methyl silicone capillary column was used
to separate and determine reactants and products. The column was
temperature programmed to start at 50 C with a 10 C/min ramp
◦
◦
◦
to reach 250 C and held for 6 min. Helium (≥99.998%, Linde) was
used as the carrier gas with a total flow rate of 30 mL/min. All the
Table 3
Properties of tested zeolite and aluminas.
H-BEA-25 Acid ␥-Al2O3
Basic ␥-Al2O3
SiO2/Al2O3
Brönsted acid sites (meq H /g)
25
1.2
+
0.46
−
Brönsted basic sites (meq OH /g)
0.57
139
0.265
62
2
SBET (m /g)
503
0.663
108
151
0.255
57
3
Pore volume (cm /g)
Mean pore diameter ( A˚ )
Skeletal density (g/cm )
3
2.237
3.281
0.105
3.327
0.105
dp (mm)
0.0081
Fig. 2. ISEC pattern for used gel-type resins.

2
4
J. Guilera et al. / Applied Catalysis A: General 413–414 (2012) 21–29
Table 4
Conversion and yield over some acid and basic catalysts at 6 h.
EOE
DEC
EOE
DEC
◦
Type
Basic
Catalyst
XDEC (%)
S
(%)
Y
(%)
T ( C)
Resin
A 26
A 21
␥-Al2O3
10.2
11.1
7.4
–
–
2.5
–
–
0.2
100
100
150
Alumina
Resin
Acid
CT224
91.8
59.7
22.8
6.2
34.0
31.4
52.8
3.5
31.2
18.8
12.0
0.2
150
150
150
150
Nafion 50
H-BEA-25
␥-Al2O3
Zeolite
Alumina
RDEC/1-octanol = 0.5, W = 2 g, 500 rpm.
species were identified by a second chromatograph equipped with
a mass spectrometer (Agilent GC/MS 5973) and chemical database
software.
be noted that molecular size of such alcohols and ethers is similar
to that of DEC and EOE, respectively. Catalyst injection was taken as
◦
zero time. Temperature was set at 100 C for basic resins (because
◦
of their low thermal stability) and 150 C for all the other catalysts.
In each experiment, DEC conversion (XDEC), selectivity to EOE
) and yield to EOE with respect to DEC (YDEC ) were estimated
by Eqs. (3)–(5), respectively.
2
.2.2. Procedure
Amberlyst 21 was previously converted to OH form and both
basic resins were first dried by methanol percolation [34], then in
EOE
DEC
EOE
−
(S
◦
◦
an oven at 80 C for 3 h at atmospheric pressure and finally at 80 C
under vacuum overnight. Acid resins were dried at 110 C for 3 h
at atmospheric pressure and subsequently at 110 C under vacuum
overnight. The residual water content of dried resins was deter-
mined by a Karl Fisher titrator (Orion AF8). Analytical volumetric
titrations showed <3% (w/w) of residual water in the tested resins,
in agreement with reported data [35,36]. Otherwise, H-BEA-25 was
activated at 500 C and both aluminas were treated at 300 C in
an atmospheric oven; subsequently dried at 110 C under vacuum
mole of DEC reacted
XDEC =
◦
× 100 [%, mol/mol]
(3)
(4)
mole of DEC initially
◦
mole of DEC reacted to form EOE
=
mole of DEC reacted
EOE
DEC
S
×
100 [%, mol/mol]
◦
◦
◦
mole of DEC reacted to form EOE
× 100
mole of DEC initially
EOE
DEC
Y
=
overnight.
The reactor was loaded with 70 mL of DEC/1-octanol mixture
EOE
DEC
XDEC × S
(
1:2 molar ratio), heated up to the desired temperature and stirred
=
[%, mol/mol]
(5)
at 500 rpm. Pressure was set at 25 bar with N to maintain the liquid
100
2
phase. When the mixture reached the working temperature, 2 g
of dried catalyst were injected into the reactor from an external
Experiments performed on Amberlyst 21, Amberlyst 26 OH, H-
cylinder by shifting with N . Resins were used with the commercial
BEA-25, Amberlyst 15, Amberlyst 35, Amberlyst 48, Amberlyst 46,
Amberlyst 70, CT224 and Dowex 50Wx2-100 were replicated three
times to assure the reproducibility of the results. Thus, data shown
in this study has a relative experimental error lower than 1.2% for
2
distribution of particles sizes, and zeolite and aluminas as a powder.
Working conditions were selected since, as quoted in literature,
liquid phase reactions of dehydration of 1-pentanol and 1-hexanol
to linear ethers take place at these conditions in the same set-up
free of external and internal mass transfer influences [10,37]. It is to
EOE
DEC
EOE
XDEC, 2.0% for S
and 2.6% for YDEC , (95% confidence level) at 8 h
of reaction time.
◦
Fig. 3. Product distribution profile in liquid phase along time over CT224. T = 150 C, RDEC/1-octanol = 0.5, W = 2 g, 500 rpm. (ꢀ) Ethanol, (♦) EOC, (ꢁ) EOE, (᭹) DEE, (ꢂ) DNOE. The
error bars indicate the confidence interval at a 95% probability level.

J. Guilera et al. / Applied Catalysis A: General 413–414 (2012) 21–29
25
◦
Fig. 4. Product distribution profile in liquid phase along time over H-BEA-25. T = 150 C, RDEC/1-octanol = 0.5, W = 2 g, 500 rpm. (ꢀ) Ethanol, (♦) EOC, (ꢁ) EOE, (᭹) DEE, (ꢂ) DNOE.
The error bars indicate the confidence interval at a 95% probability level.
3
. Results and discussion
it is to be mentioned that zeolites have been reported as suitable
catalyst to produce linear fuel ethers because they proved to be less
sensitive to water inhibition, even at higher temperature [39].
On the other hand, it is observed in Table 4 that H-BEA-25 is less
active but more selective to EOE than CT224. Figs. 3 and 4 show the
liquid product distribution along time over CT224 and H-BEA-25.
In addition to EOE, DNOE, DEE, ethanol and EOC, other by-products
were detected in low amount (<3% molar): water, dioctyl carbonate
(DOC), olefins and branched ethers, not shown for the sake of clar-
ity. As seen, the product distribution profiles showed quite different
trends along time. Over H-BEA-25, all products increased its molar
percentage with time and EOC was low along the experiment. On
the contrary, on CT224 a maximum in the concentration profile for
ethanol and EOC was observed. As these products were consumed,
the moles of EOE were increased steadily. It seems that on CT224
transesterification to EOC is faster than its decomposition to EOE.
This trend is opposite to that observed on the zeolite where EOC
is decomposed as it is formed. As expected selectivity to EOE rise
steadily through the experiment.
The ethanol and EOC profiles suggest the reaction scheme of
Fig. 5, in agreement with those reported in open literature to
synthesize ethers from dimethyl carbonate [14,40]. EOE synthesis
proceeds in two consecutive steps; the first (a) is the carboxylation
of DEC to produce EOC and the second (b) the decomposition of
EOC to EOE. The carboxylation step (a) would be kinetically favored
over acid resins. However, reaction network is complex and the
following side reactions take place in the reaction system:
3.1. Preliminary experiments
◦
Blank experiments without catalyst were performed at 150 C.
The non-catalyzed reaction allowed only the carboxyethylation
of 1-octanol to ethyl octyl carbonate (EOC) but in small amounts
(
XDEC = 6% at 6 h) and always XDEC < 1% at the initial time. In addi-
tion, it is seen that decomposition of carbonates (DEC and EOC)
was not significant in the absence of catalyst. On the other hand,
decomposition of DEC to DEE was checked over Amberlyst 121.
Results state that the acid resin decomposes significantly DEC to
DEE (XDEC = 17% at 6 h).
3.2. Catalyst screening
Table 4 summarizes the results of the catalyst screening runs. In
general, the runs carried out over basic resins and both aluminas
showed that DEC reactivity was low; therefore the EOE synthesis
was not relevantly achieved. In the case of basic resins, the low
conversions were probably due to the low working temperature.
Unlike basic catalysts, DEC conversion is improved over acid cata-
lysts. Specifically, the two types of catalysts that showed a relevant
activity were acid resins and the zeolite, although at the work-
ing temperature the higher yields were achieved over acid resins.
Similarly, zeolites were found to be less active catalyst to produce
di-n-pentyl ether than acid ion-exchange resins [38]. Nevertheless,
Fig. 5. Reaction scheme of EOE synthesis from DEC and 1-octanol.

2
6
J. Guilera et al. / Applied Catalysis A: General 413–414 (2012) 21–29
Table 5
seen that in both gel-type and macroreticular resins EOE yield
Ether average composition over Dowex 50Wx2 at 48 h.
increases as the DVB% decreases, as a consequence of the higher
DEC conversions and EOE selectivities observed. By comparing
the macroreticular conventionally sulfonated resins Amberlyst 15,
Amberlyst 16 and Amberlyst 39 resins with Amberlyst 46 it is seen
that EOE yield on the last one is far smaller than on the other three
resins, as expected because of the very low number of acid sites
of Amberlyst 46. The low EOE yield is mainly due to the very low
DEC conversion since EOE selectivity of Amberlyst 46 is only a bit
Products
EOE
DEE
DNOE
%
%
(mol/mol)
(w/w)
60.5
60.8
20.2
9.5
19.3
29.6
◦
T = 150 C, RDEC/1-octanol = 0.5, W = 2 g, 500 rpm.
(
(
(
1) DEC decomposition to DEE (c), which is undesired because of
the extra consumption of diethyl carbonate.
lesser. Amberlyst 15 has only about a 5% of –SO H groups placed
3
+
2) Carboxyocthylation of EOC (g) and subsequent decomposition
to DNOE (h) as revealed by the presence of DOC as byproduct.
3) Additionally to carbonate decomposition routes, linear ethers
can be also produced from alcohols dehydration (d, e, f), because
water was detected in the mixture as by-product.
at the polymer surface (≈0.25 meq H /g) [41]. Since this quantity is
a half of the acid capacity of Amberlyst 46 (sulfonated only at the
polymer surface) it can be inferred that in macroreticular resins
other than Amberlyst 46 the reaction takes place essentially in the
gel phase. Gel phase becomes active by the swelling of the polymer
by permeation of alcohol and the retention of part of water released
in the reaction.
Since the formed ethanol was partially dehydrated either with
another ethanol (d) or with 1-octanol (e), it could be considered as
an intermediate rather than a final product. As a result, the initial
molar ratio (DEC/1-octanol) would be an important factor to hinder
the dehydration of ethanol to DEE and, as a consequence, to increase
the selectivity to EOE [24].
Catalytic behavior of Amberlyst 15, Amberlyst 16 and Amberlyst
39 can be explained looking at the structure of the swollen gel
phase. Gel phase density of swollen resins decrease in the order
−
2
Amberlyst 15 (2 nm ; spaces between chains are equivalent to
−
2
pores of less than 1 nm diameter [42]) > Amberlyst 16 (1.5 nm
spaces are equivalent to pores of 1 nm diameter [42]) > Amberlyst
;
−
2
3
[
3
9 (0.8 nm ; spaces are equivalent to pores of 1.5 nm diameter
42]). The number of accessible sites to the reaction in Amberlyst
9 is higher, and spaces between polymer chains allow octanol and
3.3. Acid resins
As the catalyst screening revealed that the highest EOE yield was
DEC to permeate and to accommodate better reaction intermedi-
ates. As a consequence, a higher DEC conversion and EOE selectivity
is observed on Amberlyst 39. Similarly, Amberlyst 36 gives a better
EOE yield than Amberlyst 35.
achieved over acid styrene-DVB resins on a catalyst weight basis,
3 resins of different morphology and acid capacity were tested.
1
The product distribution trends for all acid resin runs were similar
to that shown in Fig. 3 and at 8 h almost all DEC was converted to
ethylated products. However, there was still a significant amount
of the intermediate products (EOC and ethanol), which should be
able to produce more EOE. As a result, to study the potential of the
system, a long time experiment was carried out on Dowex 50Wx2.
It was seen that at 48 h all intermediates were consumed. Table 5
shows the final distribution of ethers: EOE was the main ether fol-
lowed by DEE and DNOE, on a molar basis. Some octenes were also
formed but ethene was no detected.
In the case of gel-type resins the higher EOE yield obtained on
Amberlyst 121 and Dowex 50Wx2 are due to the fact that they show
the higher Vsp values. In this way they are more swollen in the reac-
−
2
tion medium and, in addition, with a polymer density of 0.4 nm
equivalent to pores of 2.6 nm [42]) they give a better selectivity
to EOE. Finally, by comparing data of Amberlyst 39 and Amberlyst
0, both resins conventionally sulfonated, it is seen that the latter
(
7
shows a better EOE yield in spite of its acid capacity is about 60% of
that of Amberlyst 39. Although they have similar swollen gel-phase
Table 6 shows DEC conversion, EOE selectivity and yield for
tested acid styrene-DVB resins after 8 h. As a general rule gel type
resins show a better behavior than macroreticular ones. It is also
−
2
volume, Amberlyst 70 has lower polymer density (0.4 nm which
is equivalent to pores of 2.6 nm diameter [42]) than Amberlyst 39.
Fig. 6. Influence of resin acidity on yield to EOE with respect to DEC (Y DE EO CE ) at 8 h. T = 150 C, RDEC/1-octanol = 0.5, W = 2 g, 500 rpm. (ꢃ) Macroreticular, (ꢂ) gel-type, (ꢄ) Amberlyst
◦
4
6. The error bars indicate the confidence interval at a 95% probability level.

J. Guilera et al. / Applied Catalysis A: General 413–414 (2012) 21–29
27
Fig. 7. Influence of swollen resin volume on yield to EOE with respect to DEC (Y DE EO CE ) at 8 h. T = 150 C, RDEC/1-octanol = 0.5, W = 2 g, 500 rpm. (ꢃ) Macroreticular, (ꢂ) gel-type, (ꢄ)
◦
Amberlyst 46. The error bars indicate the confidence interval at a 95% probability level.
Again, the wider spaces between polymer chains favored the dif-
fusion of 1-octanol and EOE as well as less steric restriction for the
reaction intermediate.
value. Likewise, the pair Amberlyst 16/Amberlyst 36 (macroretic-
ular, 12% DVB) shows better DEC conversion on Amberlyst 36 (in
line with its higher acid capacity) but a higher EOE selectivity on
Amberlyst 16 (in line with its higher Vsp value); in this case spaces
between chains are almost equivalent). However, by comparing the
pair CT124/CT224 (gel-type, 4% DVB) it is seen that the second has
better DEC conversion (in line with the higher acid capacity) but it
also shows a better EOE selectivity (despite its Vsp value is smaller,
as well as the spaces between chains). It seems again that the effect
of polymer swelling is more significant than the fact of having more
acid centers if they are not accesible.
However, the effect of the sulfonation degree is not clear because
does not always upgrade the EOE yield (see Table 6). For instance,
comparing among macroreticular resins with high percentage of
DVB, Amberlyst 48/Amberlyst35/Amberlyst 15, it is seen that the
best EOE yield is obtained on Amberlyst 48. It has the highest acid
capacity of the three resins and shows the higher DEC conversion.
In addition, despite of the fact that it is the least swollen of the
three, EOE selectivity is something higher. Probably because the
swollen gel phase in Amberlyst 48 has a smaller gel phase den-
sity (equivalent to 1 nm pores). EOE yield on Amberlyst 35 is worse
than on Amberlyst 15 probably because the fraction of accessible
acid sites is smaller in the former as pointed out by its lower Vsp
In order to state the influence of the resin structure on the cat-
alytic behavior, the effect of acid capacity and structural parameters
in swollen state on EOE yield was studied. Firstly, Fig. 6 plots the EOE
yield versus the acid capacity. As shown, acid capacities of tested
Fig. 8. Influence of DVB resin content on yield to EOE with respect to DEC (Y DE EO CE ) at 8 h. T = 150 C, RDEC/1-octanol = 0.5, W = 2 g, 500 rpm. (ꢃ) Macroreticular, (ꢂ) gel-type. The
◦
error bars indicate the confidence interval at a 95% probability level.

2
8
J. Guilera et al. / Applied Catalysis A: General 413–414 (2012) 21–29
Fig. 9. Variation of the yield to EOE with respect to DEC (Y DE EO CE ) vs. the acid centers per volume unit in a swollen polymer state (H /Vsp) at 8 h. T = 150 C, RDEC/1-octanol = 0.5,
+
◦
W = 2 g, 500 rpm. (ꢃ) Macroreticular, (ꢂ) gel-type, (ꢄ) Amberlyst 46. The error bars indicate the confidence interval at a 95% probability level.
Table 6
XDEC/S
EOE
DEC
EOE
DEC
◦
(%) at 8 h of acid resins related to their structure type, DVB content and sulfonation degree. T = 150 C, RDEC/1-octanol = 0.5, W = 2 g, 500 rpm.
/Y
Type
DVB (%)
High
Conventional Sulfonated
79.3/27.4/21.8a
Over-sulfonated
Surface Sulfonated
24.8/23.1/5.7d
Others
69.8/27.0/18.8b
Macroreticular
Gel-type
c
8
3.7/29.03/24.3
85.4/32.5/27.8e
91.7/29.6/26.9
f
1
8
2
g
83.6/38.1/31.8h
93.6/32.6/30.5
8
4
2
90.9/33.2/30.2i
j
95.0/34.9/33.1k
91.7/32.1/29.5
l
96.6/34.3/33.2
m
9
5.8/34.6/33.2
a
Amberlyst 15, bAmberlyst 35, cAmberlyst 48, dAmberlyst 46, eAmberlyst 16, Amberlyst 36, gAmberlyst 39, hAmberlyst 70, Dowex 50Wx8, CT124, CT224, Dowex 50Wx2,
Amberlyst 121.
f
i
j
k
l
m
resins were similar (except for the particular resins Amberlyst 46
and Amberlyst 70) but they showed quite different activities in
terms of EOE yield. It is also seen that gel-type resins gave higher
yields than macroreticular ones at the same acid capacity, high-
lighting the significant role of the resin structure.
As mentioned above, Vsp (specific volume of swollen polymer)
is a parameter that allows to know how much the resin swells
in the reaction medium. Fig. 7 shows the positive effect of Vsp
synthesis of EOE. It is seen that measured EOE yield decreases on
increasing [H ]/Vsp. It is to be noted that neither Amberlyst 46 (with
+
all acid sites in the polymer surface) nor Amberlyst 48 (the resin
with the highest acid capacity but also the less swollen one) follow
the general trend. It can be concluded that the higher EOE yields are
given by resins with low density of acid centers in the swollen poly-
+
3
mer volume (less than 3 meq H /cm ) and acid capacities of about
+
5 meq H /g or a bit higher. These requirements would be fulfilled
EOE
on Y
. It is seen that measured EOE yield increases with Vsp
by ion-exchange resins with high Vsp values and preferably con-
ventionally sulfonated. Between tested resins, Dowex 50Wx2-100,
Amberlyst 121 and CT224 show the higher EOE yields.
DEC
3
until a plateau is reached for Vsp values of 2 cm /g. As shown in
Figs. 1 and 2, both in macroreticular and in gel-type resins as
Vsp increases density of polymer gel phase decreases. As a result,
gel-phase is flexible enough and it could accommodate better the
reaction intermediates, and the higher space between the polymer
chains allows large molecules such as 1-octanol to access easier to
a larger number of acid centers. Figs. 1 and 2 also show that resins
with low DVB content have lower polymer fraction density in a
4. Conclusions
The catalyst screening revealed that EOE can be successfully pro-
duced in liquid phase from DEC and 1-octanol over acid catalysts
◦
at 150 C. High DEC conversion and high EOE yield were achieved
over acid resins.
EOE
polar medium and higher Vsp. Fig. 8 illustrates that YDEC increased
A two-step pathway for EOE synthesis is proposed. Firstly, the
transesterification of DEC to EOC takes place. Subsequently, EOC
decomposes to EOE. Unfortunately, direct decomposition of DEC
to DEE also occurs. Besides carbonate decomposition route, linear
ethers are also produced from alcohols dehydration reactions.
On acid resins, after 8 h of reaction almost all DEC reacted to
ethylated products. When the intermediates (EOC and ethanol)
were consumed at 48 h, EOE was the main reaction product,
followed by the two symmetrical ethers DEE and DNOE, on a molar
basis.
on decreasing the DVB resin content. Accordingly, in order to obtain
efficiently large molecules such as EOE, ion-exchange resins with
3
Vsp of 2 cm /g or higher which correspond to resins with less
than 8% of DVB are the most suitable. Between tested resins gel-
type Amberlyst 121, CT-224 and Dowex 50Wx2-100 fulfill such
requirements.
Finally, a parameter that estimates the concentration of acid
+
centers per volume unit in swollen polymer state is the [H ]/Vsp
+
3
+
ratio (meq H /cm ). Fig. 9 shows the influence of [H ]/Vsp on the

J. Guilera et al. / Applied Catalysis A: General 413–414 (2012) 21–29
29
The synthesis of EOE is highly related to structural resins prop-
erties. The accessibility of large molecules to acid centers is favored
over resins with large space between polymer chains. Conse-
quently, in order to synthesize large ethers such as EOE, a greatly
expanded polymer network in swollen state is the most suitable
resin property. It is also desirable that density of acid centers in the
swollen resin would be low. These requirements can be found in
low DVB content resins (e.g., gel-type resins as Dowex 50Wx2-100,
Amberlyst 121 or CT224).
[15] P. Tundo, F. Trotta, G. Moraglio, F. Ligorati, Ind. Eng. Chem. Res. 27 (1988)
565–1571.
16] B.C. Dunn, C. Guenneau, S.A. Hilton, J. Pahnke, E.M. Eyring, Energy Fuels 16
2002) 177–181.
[17] N. Roh, B.C. Dunn, E.M. Eyring, R.J. Pugmire, H. Meuzelaar, Fuel Proc. Technol.
3 (2003) 27–38.
1
[
(
8
[
[
18] T. Sakakura, K. Kohno, Chem. Commun. (2009) 1312–1330.
19] N. Keller, G. Rebmann, V. Keller, J. Mol. Catal. A: Chem. 317 (2010) 1–18.
[20] G. Centi, S. Perathoner, Catal. Today 148 (2009) 191–205.
[21] H. Kurakata, Y. Izumi, K. Aika, Chem. Commun. (1996) 389–390.
[
[
[
22] Y. Izumi, Platinum Met. Rev. 41 (1997) 166–170.
23] T. Sakakura, J. Choi, H. Yasuda, Chem. Rev. 107 (2007) 2365–2387.
24] S. Pros, Ms. Ch. Eng. Thesis, University of Barcelona, 2009.
Acknowledgements
[25] Dowex website, www.dow.com (04/2011).
26] J. Tejero, F. Cunill, M. Iborra, J.F. Izquierdo, C. Fité, J. Mol. Catal. A: Chem. 182–183
2002) 541–554.
[27] K. Jerabek, ACS Symp. Ser. 635 (1996) 211–224.
[28] A.G. Ogston, Trans. Faraday Soc. 54 (1958) 1754–1757.
[
(
Financial support was provided by the State Education, Uni-
versities, Research and Development Office of Spain (project
CTQ2010-16047). The authors thank Rohm and Haas France and
Purolite for providing Amberlyst and CT ion-exchange resins,
respectively. We also thank Dr. Karel Jerabek of Institute of
Chemical process Fundamentals (Prague, Czech Republic) for the
morphological analyses made by the ISEC method.
[
29] K. Jerabek, Anal. Chem. 57 (1985) 1598–1602.
[
30] A. Biffis, B. Corain, M. Zecca, C. Corvaja, K. Jerabek, J. Am. Chem. Soc. 117 (1995)
1603–1606.
[31] D.C. Sherrington, Syntheses and Separations using Functional Polymers, John
Wiley & Sons, Essex, 1998.
[
[
[
32] P.F. Siril, H.E. Cross, D.R. Brown, J. Mol. Catal. A: Chem. 279 (2008) 63–68.
33] K.D. Topp, Dow Chemical, personal communication, 2011.
TM
34] Rohm and Haas Company, Lab. Proc. for Testing Amberlyst
2004).
Polym. Catal.
(
References
[
[
35] M. Iborra, C. Fité, J. Tejero, F. Cunill, J.F. Izquierdo, React. Polym. 21 (1993) 65–76.
36] M. Iborra, J. Tejero, F. Cunill, J.F. Izquierdo, C. Fité, Ind. Eng. Chem. Res. 39 (2000)
[
[
[
[
[
[
[
1] G.A. Olah, US Patent WO 95/25153 (1995).
2] Directive 2009/30/EC (2009).
3] S. Rossini, Catal. Today 4 (2003) 467–484.
4] M. Kozak, J. Merkisz, Kom. Mot. Energ. Roln. TEKA 7 (2007) 129–136.
5] C.Y. Choi, R.D. Reitz, Fuel 78 (1999) 1303–1317.
1416–1422.
[
37] E. Medina, R. Bringué, J. Tejero, M. Iborra, C. Fité, J.F. Izquierdo, F. Cunill, Proceed-
ings of 6th European Congress of Chemical Engineering, Copenhagen, 2007,
www.nt.ntnu.no/users/skoge/prost/proceedings/ecce6 sep07/upload/1686.pdf.
38] J. Tejero, C. Fité, M. Iborra, J. Izquierdo, F. Cunill, R. Bringué, Microporous Meso-
porous Mater. 117 (2009) 650–660.
[
6] C. Beatrice, C. Bertoli, N.D. Giacomo, Comb. Sci. Technol. 137 (1998) 31–50.
7] M. Kozak, J. Merkisz, P. Bielaczyc, A. Szczotka, SAE Technical Paper 2009-01-
[39] R. Nel, A. Klerk, Ind. Eng. Chem. Res. 48 (2009) 5230–5238.
[40] P. Tundo, F. Arico, A. Rosamilia, E. Memoli, Green Chem. 10 (2008) 1182–1189.
[41] S. Ihm, M. Chung, K. Park, Ind. Eng. Chem. Res. 27 (1988) 41–45.
[42] K. Jerabek, L. Holub, 13th International IUPAC Conference on Polymers and
Organic Chemistry, Montreal, 2009.
2
696, 2009.
[8] G.C. Pecci, M.G. Clerici, F. Giavazzi, F. Ancillotti, M. Marchionna, R. Patrini, Proc.
IX Int. Symp. Alcohol Fuels 1 (1991) 321–326.
[9] A.W. Williamson, J. Chem. Soc. 4 (1852) 229–239.
[
[
[
10] R. Bringué, M. Iborra, J. Tejero, J.F. Izquierdo, F. Cunill, C. Fité, V.J. Cruz, J. Catal.
44 (2006) 33–42.
11] E. Medina, R. Bringué, J. Tejero, M. Iborra, C. Fité, Appl. Catal. A: Gen. 374 (2010)
1–47.
12] C. Casas, R. Bringué, E. Ramírez, M. Iborra, J. Tejero, Appl. Catal. A: Gen. 396
2011) 129–139.
[
43] M. Girolamo, M. Lami, M. Marchionna, D. Sanflippo, M. Andreoni, A. Galletti, G.
Sbrana, Catal. Lett. 1–2 (1996) 127–131.
2
[
[
44] A. Bellamy, React. Polym. 23 (1994) 101–106.
45] P.K. Kiviranta-Pääkkönen, L.K. Struckmann (née Rihko), J.A. Linnekoski, A.O.I.
Krause, Ind. Eng. Chem. Res. 37 (1998) 18–24.
4
(
[
46] M. Honkela, A. Root, M. Lindblad, A. Krause, Appl. Catal. A: Gen. 295 (2005)
[
[
13] P. Tundo, L. Rossi, A. Loris, J. Org. Chem. 70 (2005) 2219–2224.
14] P.N. Gooden, R.A. Bourne, A.J. Parrott, H.S. Bevinakatti, D.J. Irvine, M. Poliakoff,
Org. Process Res. Dev. 14 (2010) 411–416.
216–223.
[
47] D. Kapila, US Patent 2005/0137428 A1 (2005).